Competitive Binding of Protein Kinase Cα to ... - ACS Publications

Nov 9, 2006 - Anthony C. Cook, Cojen Ho, Jennifer L. Kershner, Steve A. Malinowski, Heath Moldveen, Brigid A. Stagliano, and. Simon J. Slater*. Depart...
0 downloads 0 Views 254KB Size
14452

Biochemistry 2006, 45, 14452-14465

Competitive Binding of Protein Kinase CR to Membranes and Rho GTPases† Anthony C. Cook, Cojen Ho, Jennifer L. Kershner, Steve A. Malinowski, Heath Moldveen, Brigid A. Stagliano, and Simon J. Slater* Department of Pathology and Cell Biology, Thomas Jefferson UniVersity, Philadelphia, PennsylVania 19107 ReceiVed June 22, 2006; ReVised Manuscript ReceiVed October 2, 2006

ABSTRACT:

Previously, we have shown that protein kinase CR (PKCR) forms a direct high-affinity, isozymespecific and membrane lipid-independent interaction with Rho GTPases [Slater, S. J., Seiz, J. L., Stagliano, B. A., and Stubbs, C. D. (2001) Biochemistry 40, 4437-4445]. Since the cellular activation of PKCR involves an initial translocation from cytosolic to membrane compartments, the present study investigates the interdependence between the direct protein-protein interaction of PKCR with the Rho GTPase, Cdc42, and the protein-lipid interactions of PKCR with membranes. It was hypothesized that the interaction of PKCR with membrane-bound Cdc42 would contribute to the overall membrane-binding affinity of the kinase by providing an additional anchor. However, it was found that the incorporation of isoprenylated Cdc42 into membranes resulted in an apparent decrease in the membrane-binding affinity of PKCR, whereas the association of PKCβI, PKCδ, PKC, and PKCζ was each unaffected. The presence of membranebound Cdc42 resulted in a rightward shift in both the PS- and Ca2+-concentration response curves for PKCR membrane association and for the ensuing activation, whereas the maximal levels of binding and activation attained at saturating PS and Ca2+ concentrations were in each case unaffected. Overall, these findings suggest that PKCR undergoes a isozyme-specific interaction with membrane-bound Cdc42 to form a PKCR-Cdc42 complex, which possesses a membrane-binding affinity that is reduced relative to that of the individual components due to competition between Cdc42 and PS/Ca2+ for binding to PKCR. Consistent with this, it was found that the interaction of PKCR with membrane-bound Cdc42 was accompanied by the physical dissociation of the PKCR-Cdc42 complex from membranes. Thus, the study provides a novel mechanism by which the membrane association and activation of PKCR and Cdc42 may be regulated by competing protein-protein and protein-lipid interactions.

Protein kinase C (PKC)1 consists of a family of minimally 12 serine-threonine kinases that occupy critical nodes in receptor-initiated signaling networks which regulate both normal and pathological cellular processes, including secretion, proliferation, differentiation, apoptosis, and migration (1-8). Each of the isozymes can be classified into three groups based upon the presence or absence of structurally conserved domains that dictate activator dependences for membrane association and activation (5, 7, 9). In the case of the “conventional” PKCR, -βI/βII, and -γ isozymes, these include the activator-binding C1 domains and the Ca2+binding C2 domain. The C1 domains consist of a tandem C1A and C1B arrangement, each of which can potentially † This work was supported by U.S. Public Health Service Grants AA010990 and AA010968. * To whom correspondence should be addressed. Tel: (215) 5035019. Fax: (215) 923-2218. E-mail: [email protected]. 1 Abbreviations: DOGS-NTA/Ni2+, 1,2-dioleoyl-sn-glycero-3-{[N(5-amino-1-carboxypentyl)iminodiacetic acid]succinyl} nickel salt; EYPC, chicken egg L-R-phosphatidylcholine; FC, flow cell; FRET, fluorescence resonance energy transfer; GMP-PNP, guanosine 5′-[β,γimido]triphosphate trisodium salt; HAF, 5-hexadecanoylaminofluorescein; HSC-T6, immortalized hepatic stellate cell line-T6; LUV, large unilamellar vesicles; MBP4-14, bovine myelin basic protein peptide substrate; MANT-GMP-PNP, 3′-O-(N-methylanthraniloyl)-β,γ-imidoguanosine 5′-triphosphate trisodium salt; PKCR, R-isoform of protein kinase C; POPS, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoserine; RhoGDIR, R-isoform of Rho GDP-dissociation inhibitor; SPR, surface plasmon resonance.

bind the endogenous activator, diacylglycerol, and exogenous activators including phorbol esters (10, 11). The “novel” PKCδ, -, -η, -θ, and -µ isozymes contain C2 domains that lack Ca2+-binding ability, while retaining functional C1A and C1B domains. The “atypical” PKCζ, -ι, and -λ regulatory domains also lack a functional C2 domain and contain a single C1 domain that lacks the ability to bind activators, the function of which remains obscure. Each isozyme becomes catalytically competent by undergoing multiple serine-threonine and tyrosine phosphorylations that are either autocatalytic or catalyzed by another upstream kinase, such as the phosphoinositide-dependent kinase, PDK1 (6). The cellular activation of PKC isozymes is accompanied by a reversible translocation from cytosol to membrane compartments that is elicited by receptor- and phospholipase C-coupled DAG production and inositol trisphosphatemediated Ca2+ release from internal stores (5). The initial interaction of the conventional PKC isozymes with the membrane is mediated by the formation of a relatively low affinity Ca2+ bridge between anionic residues in the C2 domain and the head group of phosphatidylserine (PS), which then facilitates the interaction of either DAG with the C1A domain or the tumor-promoting phorbol esters with the C1B domain (5, 11, 12). These parallel but independent proteinlipid interactions supply the activation energy for a conformational change in the membrane-associated PKC molecule

10.1021/bi0612420 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/09/2006

Interaction of PKCR with Rho GTPases and Membranes that leads to the displacement of an autoinhibitory pseudosubstrate from the active site, allowing substrate binding and phosphorylation to occur (13-15). The low molecular weight Rho GTPases that belong to the Ras superfamily each play a central role in many cellular processes, including the regulation of actin cytoskeleton dynamics, gene transcription, cell cycle progression, and membrane trafficking (16, 17). The closely related mammalian Rho GTPase family consists of minimally 14 members, of which RhoA, Cdc42, and Rac1 have been widely studied with respect to their roles in the regulation of the cytoskeleton (18, 19), and their structures and molecular mechanisms of action have been characterized in detail (16, 20). Similar to the PKC isozymes, the Rho GTPases also act as tightly regulated molecular switches by cycling between inactive GDP-bound cytosolic forms and active GTP-bound membrane-associated forms. Each of the Rho GTPases are retained in an inactive GDP-bound state in the cytosol by interaction with the Rho GDP-dissociation inhibitor (RhoGDI), and translocation to the membrane is triggered by a guanine nucleotide exchange factor (GEF) catalyzed GDP exchange for GTP (16, 21). Membrane association is also coupled to posttranslational modifications which consist of the sequential attachment of a geranylgeranyl chain to a cysteine residue within a conserved C-terminal CAAX domain, followed by proteolysis of the three Cterminal residues and carboxymethylation (20). In general, the GTP-bound membrane-associated form of each Rho GTPase can then specifically interact with and activate membrane-associated downstream targets or effectors, examples of which include phospholipases D (22-24) and C (25, 26), rhophillin (27), diacylglycerol kinase θ (28),several protein kinases including PKN (27, 29), PRK1, and 2 (30, 31), and the Rho-associated kinases p160-ROCK and p150ROK-R and -β (32, 33). Recent studies have provided evidence supporting crosstalk between PKC and Rho GTPase mediated signaling pathways and have revealed a close association between the mammalian proteins (34-41) and also between the yeast homologues Pkc1p and Rho1p (42, 43). In particular, studies from this laboratory have shown that PKCR undergoes a direct, isozyme-specific and lipid-independent proteinprotein interaction with RhoA, Cdc42, and to a lesser extent Rac1 (44, 45). In each case this interaction was shown to result in kinase activation with respect to the phosphorylation of a downstream PKCR substrate, supporting the notion that PKCR should be added to the list of downstream effectors for these Rho GTPases. Both the interaction and the ensuing PKCR activation were found to be dependent on the Ca2+and DAG/phorbol ester-induced conformational state of PKCR and also on the GTP/GDP-induced conformational states of the Rho GTPases. Whereas it was necessary to exclude membrane lipids from the assay systems used in the previous study in order to provide evidence for a direct protein-protein interaction between PKCR and Rho GTPases (44), the conformational changes in each protein that result in activation both occur at the membrane. Thus, an important question remains whether there is interdependence between the direct protein-protein interaction of PKCR with Rho GTPases and the protein-lipid interactions that mediate the membrane association of the kinase.

Biochemistry, Vol. 45, No. 48, 2006 14453 The aim of this study was to investigate the interplay between protein-lipid and protein-protein interactions in the association of PKCR with membranes containing the Rho GTPase, Cdc42. It was initially hypothesized that if the direct PKCR-Cdc42 interaction occurred at the membrane surface, then this would result in an enhancement of PKCR membrane association by providing an additional anchoring element in addition to Ca2+/PS binding to the C2 domain and DAG/ phorbol ester binding to the C1 domains. However, it was found that the incorporation of Cdc42 into membranes resulted in a decrease in the membrane-binding affinity of PKCR, which was observed as a rightward shift in the PSand Ca2+-concentration dependencies for both membrane association and activation. Thus, whereas the results indicate that the previously reported direct interaction between PKCR and Cdc42 also occurs at the membrane surface, the findings suggest that the binding affinity of the resultant PKCRCdc42 complex is reduced relative to each of the individual protein components. Consistent with this, the PKCR-Cdc42 complex formed by the direct interaction between the two proteins at the membrane surface was found to physically dissociate from the membrane. These findings provide a novel mechanism by which the interaction of PKCR and Cdc42 with downstream targets could be regulated by competing protein-protein and protein-lipid interactions. MATERIALS AND METHODS Materials. Primers, Grace’s insect media, phosphatebuffered saline (PBS), 5-hexadecanoylaminofluorescein (HAF), 3′-O-(N-methylanthraniloyl)-β,γ-imidoguanosine 5′-triphosphate trisodium salt (MANT-GMP-PNP), and Spodoptera frugiperda (Sf9) cells were each obtained from Invitrogen Technologies (Carlsbad, CA), and the hepatic stellate cell line-T6 (HSC-T6) was a kind gift from Dr. William S. Blaner (Department of Medicine, Columbia University, New York). The L1 sensor chip was purchased for use in a Biacore 2000 SPR system from Biacore, Inc. (Piscataway, NJ). Fetal bovine serum (FBS), protease inhibitor cocktail, guanosine 5′-[β,γ-imido]triphosphate trisodium salt (GMP-PNP), and all other research grade chemicals were from Sigma (St. Louis, MO). Chicken egg L-R-phosphatidylcholine (EYPC), 1-palmitoyl-2-oleoylphosphatidylserine (POPS), and 1,2dioleoyl-sn-glycero-3-{[N-(5-amino-1-carboxypentyl)iminodiacetic acid]succinyl} nickel salt) (DOGS-NTA/Ni2+) were from Avanti Polar Lipids, Inc. (Alabaster, AL). Adenosine 5′-triphosphate (ATP) was from Boehringer Mannheim (Indianapolis, IN), and [γ-32P]ATP was from New England Nuclear (Boston, MA). Baculovirus encoding human Cdc42 N-terminal tagged with 6×His was a kind gift from Dr. R. A. Cerione (Department of Molecular Medicine, Veterinary Medical Center, Cornell University, New York), and a baculovirus containing RhoGDI N-terminal tagged with GST was custom manufactured by Orbigen (San Diego, CA). Where required, free Ca2+ and Mg2+ concentrations in EDTA buffers were calculated using the program “MAXC” at http:// www.stanford.edu/∼cpatton/maxc.html (46), which takes into account the total concentrations of EDTA, Ca2+, and Mg2+ present in the assay system. BaculoVirus Construction. Recombinant PKCR, PKCβI, PKCδ, PKC, and PKCζ (rat brain) were each prepared using the Bac-to-Bac baculovirus Sf9 cell expression system (Invitrogen, Carlsbad, CA) as originally described (47), with

14454 Biochemistry, Vol. 45, No. 48, 2006 modifications (48), and purified to homogeneity by following published procedures (48, 49). The isoforms PKCδ, PKC, and PKCζ were overexpressed in Sf9 cells as fusion proteins containing a 6×His attached to the C-terminus (48) and were purified as described previously (48, 50). Baculovirus containing human RhoA N-terminal tagged with 6×His was also prepared using the Bac-to-Bac baculovirus SF9 cell expression system (Invitrogen, Carlsbad, CA). Briefly, the RhoA coding sequence was double digested from pUSEampRhoA (Upstate, Lake Placid, NY) using EcoRI/XhoI and ligated into the baculovirus transfer vector pfastbacHta. Correct ligation was confirmed by colony PCR, and the construct was recombined with bacmid according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA). Expression of proteins was confirmed by SDS-PAGE followed by Western blotting using the ECL system (Amersham Biosciences, Piscataway, NJ) and the appropriate primary antibodies (Cytoskeleton, Inc., Ann Arbor, MI). Coexpression of RhoA and Cdc42 with RhoGDIR. Since the presence of the geranylgeranyl chain attached to the C-terminus of RhoA and Cdc42 reduces their aqueous solubility, each protein was coexpressed and purified as a complex with RhoGDIR-GST, as described previously (51). In this complex, the geranylgeranyl chain of each Rho GTPase is shielded from the aqueous environment by insertion into a hydrophobic binding pocket on RhoGDIR (52, 53). Briefly, log phase Sf9 cells in 1 L of Grace’s insect media containing 10% FBS were infected with baculovirus encoding Cdc42-6×His or RhoA-6×His and RhoGDIR-GST at a multiplicity of infection of 5, 1, and 2, respectively, followed by incubation for 4 days. The cells were then washed once with PBS, pelleted, and then lysed using a Dounce homogenizer in buffer A (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM MgCl2) containing a protease inhibitor cocktail. The cell debris was removed by centrifugation at 3500g for 15 min (4 °C), and the resultant supernatant was centrifuged at 100000g for 1 h to pellet membranes. The supernatant was incubated with Ni2+/NTA chelating resin in buffer A for 45 min at 4 °C with gentle rocking, and the resin was then packed into a 5 mL column attached to an A ¨ KTApurifier FPLC system (GE Healthcare, Piscataway, NJ). The column was washed with buffer A until the absorbance of the flow-through at 280 nm was